Cardioids at the sides of a sphere
alberto
I own two cardioids (don't have any omnis) and I was wondering if I
could have a more natural stereo image if I placed them at the sides of
a sphere 20cm in diameter. The configuration I'm fancying has the two
cardio mikes horizontal and parallel pointing forward, with the
capsules almost touching two opposite points on the equator of the
sphere. Anyone can tell what I'm going to get and if it's worth trying?
I've taken a look at the math of the acoustics of a sphere: I believe
the timing should be the same as a true sphere microphone, but the
amplitudes might be dependent on the frequency, so I might get
coloration.
Alberto
Soundhaspriority
"alberto" <albl...@yahoo.com> wrote in message
news:1165479318.4...@16g2000cwy.googlegroups.com...
out that if you place a cardioid against a boundary surface, it will not
behave as a cardioid. The wave has to be present at the sides and back of
the capsule in order for the cancellation to occur that creates the cardioid
pattern. By placing it against the boundary, you create a very complex
angular dependence, even before the sphere shape is factored in.
All pretense of a pattern independent of frequency goes away, because in
acoustic caluculations, the size of the obstacle is measured in units of the
wavelength under consideration. So you're building in even more severe
frequency dependence on angle than exists in a real-world imperfect cardioid
capsule.
Baffle-based microphones use omnis because all of the above complexity is
removed from the microphones, leaving only the effect of frequency on the
baffle itself. And this calculation has to be done only for one frequency,
because the solution is in terms of D/lambda, where D is the baffle
dimension, and lambda the wavelength.
There are many near-coincident techniques you can exploit with cardioids, of
which ORTF and close relatives are regarded as always giving at least good
results. The spacing and angular spread should provide ample opportunity for
experimentation.
Bob Morein
Dresher, PA
(215) 646-4894
alberto
> out that if you place a cardioid against a boundary surface, it will not
> behave as a cardioid. The wave has to be present at the sides and back of
> the capsule in order for the cancellation to occur that creates the cardioid
> pattern. By placing it against the boundary, you create a very complex
> angular dependence, even before the sphere shape is factored in.
Oh, I thought I could get away just computing the direction of the
pressure gradient, taking the angle between this and the axis of the
mike and looking it up in the polar pattern (published or theoretical).
In the setup I was envisioning there is air all around the mike, but
true it is in the middle of a messy soundfield, something different
from a sum of plane waves (i.e. arguably what the manufacturer sells
you a directional mike for)...
Alberto
Mike Rivers
alberto wrote:
> Oh, I thought I could get away just computing the direction of the
> pressure gradient, taking the angle between this and the axis of the
> mike and looking it up in the polar pattern (published or theoretical).
> In the setup I was envisioning there is air all around the mike, but
> true it is in the middle of a messy soundfield, something different
> from a sum of plane waves (i.e. arguably what the manufacturer sells
> you a directional mike for)...
Oh., Well, you can try anything you want. It won't work any worse if
someone tells you that it doesn't work.
What is this sphere made from, and how do you propose to put put the
two mics directly opposite? The sum of the lengths of two mics, unless
you have those with extendable capsules, is almost always greater than
20 cm, and that's without the connectors and cables.
Why not spend a couple of dollars and get a pair of electret omni
capsules and play with something more likely to geive predictable
results? Or are you trying to invent an exercise in mathematics that
won't translate well in practice?
audioae...@gmail.com
jecklin disc, schneider disc a couple of existing swiss mic
techniques
the size of the disc/sphere dictates the cutoff frequency
the surface of the disc/shere can cause comb filtering...
alberto
> two mics directly opposite? The sum of the lengths of two mics, unless
> you have those with extendable capsules, is almost always greater than
> 20 cm, and that's without the connectors and cables.
It's a white lamp made in polycarbonate, you may find such things in
gardens or on exterior walls. It's essentially a plastic shell, few
millimeters thick and hollow. The mikes will be outside the sphere much
like the figure-of-eights in this picture:
http://www.schoeps.de/E-2004/kfm360.html
> Why not spend a couple of dollars and get a pair of electret omni
> capsules and play with something more likely to geive predictable
> results?
I need a decent S/N, but I'll think about it just to hear what the
stereo image sounds like.
> Or are you trying to invent an exercise in mathematics that
> won't translate well in practice?
Ha ha good point. I like mathematics and I have really inexperienced
ears. I suppose if the math says it's gonna be perfect, the reality
check can still be disappointing, whereas if the math says I'll get
coloration, I'm sure going to get coloration... Anyway, tonight I'll
have the chance of recording a choir performance: I'll use some sort of
ORTF for the concert but I'll record the rehearsals with the sphere, so
I can make comparisons. I have no idea what the recording angle of such
a thing might be, though.
Alberto
Scott Dorsey
>I own two cardioids (don't have any omnis) and I was wondering if I
>could have a more natural stereo image if I placed them at the sides of
>a sphere 20cm in diameter.
No, you will get a huge hole in the center. The sphere is acting as a
baffle and increasing the separation. The mikes are pointed apart, so
the separation is already too high.
The reason folks use a sphere with omnis is to act as a baffle and
increase amplitude differences between channels. With cardioids, that's
the last thing you need.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
alberto
> baffle and increasing the separation. The mikes are pointed apart, so
> the separation is already too high.
I realized I haven't been clear enough. The mikes are pointed forward,
not apart; to get an idea of what I mean please see picture in
http://www.schoeps.de/E-2004/kfm360.html
Actually, if you mix 50-50 the mikes on each side of the KFM360 and
take the two mixes as L and R channel, you should get what I'm going to
get, right? (Of course, that's not what the KFM360 is meant for.)
Alberto
Mike Rivers
alberto wrote:
> It's a white lamp made in polycarbonate, you may find such things in
> gardens or on exterior walls. It's essentially a plastic shell, few
> millimeters thick and hollow. The mikes will be outside the sphere much
> like the figure-of-eights in this picture:
>
> http://www.schoeps.de/E-2004/kfm360.html
Aha! I hadn't seen that arrangement, only a Schoeps arrangement with
two omni mics mounted inside the ball so that the business end was
flush with the surface. I guess you're thinking of the cardioids
parallel and pointing forward toward the sound source?
I suppose it would act like some sort of a baffled near-coincident
arrangement. In any case, I'd be suspicious of the potential for
resonance of the globe, both of the plastic and the air inside. The
Schoeps sphere is quite solid. You might want to fill it with something
to deaden it.
> Ha ha good point. I like mathematics and I have really inexperienced
> ears. I suppose if the math says it's gonna be perfect, the reality
> check can still be disappointing, whereas if the math says I'll get
> coloration, I'm sure going to get coloration.
Oh, you'll get coloration all right, but this is a difficult thing to
quantify, or even if you come up with some numbers, to relate those
numbers to what you hear. This sounds like a good project for a
collaboration with someone who can evaluate it as a microphone setup.
The math can (and no doubt will) predict less than perfect results, but
that's no reason not to try it.
After all, some of our most useful medicines are known to be poisonous.
If we believed "the math" and wrote them off, there would be more sick
people. <g>
Scott Dorsey
>> No, you will get a huge hole in the center. The sphere is acting as a
>> baffle and increasing the separation. The mikes are pointed apart, so
>> the separation is already too high.
>
>I realized I haven't been clear enough. The mikes are pointed forward,
>not apart; to get an idea of what I mean please see picture in
>
>http://www.schoeps.de/E-2004/kfm360.html
This is the Schoeps Sphere. Read the original paper on it. It is
designed to give you amplitude differences with omni microphones.
You already HAVE amplitude differences with cardioids, if you point
them outwards. If you point them forwards and put a baffle between
them, all you do is screw the response in the middle up more.
>Actually, if you mix 50-50 the mikes on each side of the KFM360 and
>take the two mixes as L and R channel, you should get what I'm going to
>get, right? (Of course, that's not what the KFM360 is meant for.)
Why would you want to do this? What problem are you trying to solve?
alberto
> quantify, or even if you come up with some numbers, to relate those
> numbers to what you hear. This sounds like a good project for a
> collaboration with someone who can evaluate it as a microphone setup.
> The math can (and no doubt will) predict less than perfect results, but
> that's no reason not to try it.
Why not be the collaborator yourself! I'll make some recording and
share it on the internet for any good ears in rec.audio.pro to inspect.
What is an acceptable compressed format?
Alberto
alberto
>
> This is the Schoeps Sphere. Read the original paper on it. It is
> designed to give you amplitude differences with omni microphones.
I think you're talking about the KFM6
http://www.schoeps.de/E-2004/kfm6.html
The KFM360 is different in that it has two more mics, bidirectional,
hanging outside and pointing forward to the source. (With four mics it
is designed for ambisonic.) I only have the directional mics (just
cardioids not figure-8), and no omnis. My setup is not to be compared
to KFM360, I was showing the picture on the web just to make clear how
I'll arrange my mics.
> Why would you want to do this? What problem are you trying to solve?
I'm trying to have some of the sphere stereo benefit (presented in
Guenther Theile's original paper "On the naturalness..." which I did
read) with the only two mics I own, a Rode NT5 cardioid pair, with
about a factor of 20 smaller budget. Of course I'm not expecting the
same quality as Schoeps sphere KFM6, actually anything better than 20
times less quality would be a success... :-) But I'd actually use this
setup only if it was any better, under some respect, than a standard
near-coincident setup, which I can always do. I guess timing stereo
will be "natural", amplitude stereo is more or less unpredictable and
the sound will be somehow colored (if this is a problem, I wonder how
the guys at Schoeps deal with the coloration with the KFM360?). On a
plus, I'll isolate away the rear (audience) soundfield, which is
something the KFM6 doesn't.
I wasn't expecting this idea was so original...
Alberto
Mike Rivers
alberto wrote:
> Why not be the collaborator yourself! I'll make some recording and
> share it on the internet for any good ears in rec.audio.pro to inspect.
> What is an acceptable compressed format?
You aren't going to want a lot of opinions. Why not make a recording,
put it on a CD, and ask people who are interested in listening to it to
e-mail you their mailing address, then send them a CD. You'll proalby
get a handful who are really interested.
No compressed format is really satisfactory for a critical evaluation
of anything but the effect of compression. <g>
Scott Dorsey
>> Why would you want to do this? What problem are you trying to solve?
>
>I'm trying to have some of the sphere stereo benefit (presented in
>Guenther Theile's original paper "On the naturalness..." which I did
>read) with the only two mics I own, a Rode NT5 cardioid pair, with
>about a factor of 20 smaller budget.
All of that benefit, though, comes from using omnis. The whole POINT
of using the sphere is that it gives you amplitude directionality from
an omni.
You do not need additional amplitude directionality because you have
cardioids.
>Of course I'm not expecting the
>same quality as Schoeps sphere KFM6, actually anything better than 20
>times less quality would be a success... :-) But I'd actually use this
>setup only if it was any better, under some respect, than a standard
>near-coincident setup, which I can always do. I guess timing stereo
>will be "natural", amplitude stereo is more or less unpredictable and
>the sound will be somehow colored (if this is a problem, I wonder how
>the guys at Schoeps deal with the coloration with the KFM360?). On a
>plus, I'll isolate away the rear (audience) soundfield, which is
>something the KFM6 doesn't.
The KFM360 is a standard KFM6 with some additional rear pickup added for
surround ambience. It is not, as you stated earlier, an ambisonic mike.
It is a conventional sphere with surround channels added. You can basically
think of the two as being the same thing as long as you're in the stereo
domain.
The whole purpose of this system is to give you both timing and amplitude
cues while using omnis. It gives you something approaching an ORTF type
pattern, but allows you to use omni mikes. The reason people use it is
to allow them to use omni mikes.
>I wasn't expecting this idea was so original...
It's not a good idea, which is basically why people don't do it.
David Satz
> The KFM360 is a standard KFM6 with some additional rear pickup added for
> surround ambience.
Actually it's a smaller sphere for the sake of an increased
stereophonic recording angle. Jerry Bruck's thought was that the KFM
360 would probably be placed a bit closer than the KFM 6, so he wanted
a wider pickup angle. The complete system includes an expensive DSP box
which, among other things, equalizes the omni capsules that are built
into the surface of the sphere. For details, see page 11 of
http://www.schoeps.de/PDFs/SCHOEPS_surround-brochure.pdf .
Regarding boundary layer microphones with patterns other than omni, of
course this is entirely possible, and several companies sell "PZMs" or
other such constructions made with cardioid or even supercardioid
capsules. Schoeps offers a small (ca. 4" diameter) mounting plate
called the BLC, which allows any Colette capsule (on an active cable)
or CCM microphone to be used as a boundary-layer microphone regardless
of polar pattern--though using a figure-8 for this application would be
a little hard to fathom. The result is a "half" version of the original
pattern in the vertical plane, with a corresponding increase in both
sensitivity and signal-to-noise ratio.
The nifty thing about using these with a directional capsule on, say,
the lip of a theatrical stage is that the farther back a person is from
the microphone, the more on-axis he or she is likely to be, so that
differing distances are somewhat compensated for. I've made a stereo
recording with supercardioids in those mounting plates, with relatively
decent results considering the weirdness of placement on the floor in
the first place--the singer insisted on having no microphones on stands
between her and her audience, and hanging the mikes was impractical in
that venue.
Disclaimer: I write, edit and translate various texts for Schoeps as a
consultant, and am friends with several people who work there.
--best regards
Bob Cain
> Why not be the collaborator yourself! I'll make some recording and
> share it on the internet for any good ears in rec.audio.pro to inspect.
Please do. All the speculation here aside, this sounds like a great
experiment. It's basically a time of arrival based setup and the
effect of the directivity is kinda hard to intuit. You might want to
play with the angles of the cards.
> What is an acceptable compressed format?
192-212 kbps MP3 would be a good place to start.
Bob
--
"Things should be described as simply as possible, but no simpler."
A. Einstein
alberto
> of using the sphere is that it gives you amplitude directionality from
> an omni.
> The whole purpose of this system is to give you both timing and amplitude
> cues while using omnis. It gives you something approaching an ORTF type
> pattern, but allows you to use omni mikes. The reason people use it is
> to allow them to use omni mikes.
As I understand it, there is more to it then just have, with omnis,
something that you'd have more easily with directional mics: with a
sphere (or any other baffle) you get ILD and ITD dependent on
frequency, which is something that happens to real heads in real life,
making it a more "natural" stereo feeling than any near coincident
setup with no baffle. I'm after that naturalness, but don't own any
omnis.
> The KFM360 is a standard KFM6 with some additional rear pickup added for
> surround ambience. It is not, as you stated earlier, an ambisonic mike.
Correct, sorry I wrote "ambisonic" but I meant "surround" of course.
Now I answer David Satz about the issue of directional mics against a
boundary: I agree with Soundhaspriority who first raised the issue.
David wrote correctly that this is something people have done and do,
but always (with one exception I know of) against flat surfaces. Flat
reflecting surfaces trasform a soundfield made of plain waves into
another soundfield made of plane waves. By plane wave I mean an
acoustic wave with essentially flat wave front. Both free far field and
diffuse field are assumed to be made of plain waves, correct me if I'm
wrong. Close to a sphere this is no longer the case: flat wavefronts
hitting a sphere bounce back curved, and if they are picked up by a
directional mic I'm expecting some proximity effect. The exception I've
mentioned before is the KFM360: two figure-of-eights close to a sphere.
David, do you happen to know how engineers at Schoeps have evaluated
this issue?
I thank all people who have contributed to this thread so far. I'll
share some recordings with my sphere-with-cardioids setup very soon.
Alberto
alberto
the following URL:
http://www.ringaround.it/albe/ae/cardiosphere/
The samples are available both as wav and as mp3 at 224Kbps. Any
comment will be greatly appreciated.
Alberto
Peter Larsen
> You can find a first attempt to record with my sphere-plus-cardioids at
> the following URL:
> http://www.ringaround.it/albe/ae/cardiosphere/
Thank you!
> The samples are available both as wav and as mp3 at 224Kbps. Any
> comment will be greatly appreciated.
Pending, now is not the time for listening at european longitude ....
> Alberto
Peter Larsen
Peter Larsen
The concerto file appears to be corrupted after about 55 seconds, but it
provides a much better spatial view. In comparison it appears that the
reflections from the sphere cause coloration and perspective fogging.
The issue you face is that the mics appear to need to be flush mounted
or isolated from the surface reflections via some kind of funnel. The
latter would of course increase the coloration issue, but such a design
- including for good measure a furry antirefrection coating of the
sphere has been used for quite some time.
Those that like to "do audio" AND to shave the sides of their heads
would do well by listening to these examples, the difference is as I
recall it from when I got tired of male pattern baldness and shaved the
lot off. For optimum temporal resolution the ear must be surrounded, but
not covered, by an acoustically antireflectant coat, moderate hair
length suffices.
IF you want to make a sphere work you MUST add suitable EQ, whatever
that is, and it appears to me that it also most be non-reflectant. See
also the Schoeps description of their commercial product you linked to.
Already from reading that I would have abandoned to diy it, even with
flush mounted omni's it is a high tech product.
> Alberto
Peter Larsen